Information Age (IA): Capital One has consistently won awards for the way it uses technology – why do you think it has earned this recognition and what is so unique about the company's philosophy towards investment in and deployment of technology?
Catherine Doran (CD): I think the one respect in which Capital One is very different from most other companies is that, whereas most other companies had a business model and then computerised it, so the technology's an adjunct to the main business, Capital One was put together with technology right at the core of the business. Loads of companies say that and most of the time it's junk. But Capital One is definitely different in this respect.
When Capital One's founders Nigel Morris and Rich Fairbank started the company, they looked at how they might use technology differently. They thought: ‘There's got to be an opportunity here. If we use technology to understand our customers, and understand how people work, we should be able to offer products to different segments of the population with different price points and different characteristics.' They realised they could offer a broader range of products to a wider group of customers and still make a healthy profit, meanwhile undercutting the traditional players.
They had this notion of dividing up the [potential credit card] population by segmenting our data, analysing it and drawing conclusions to see what works and what doesn't work – and then using that to target market to customers. They first applied this philosophy at Signet Bank in the US. It worked so well that, after a few years, it was such as highly profitable dimension that they spun it off as Capital One. So technology is right at the heart of Capital One's genesis. Our strategy is all centred around understanding and analysing information.
IA: Despite that firm commitment to analysing information, the credit card industry is dogged by low conversion rates from direct marketing activities. From a technology perspective, how have you managed to improve customer interaction levels?
CD: Any fool can lend money, the brains is in getting people to pay it back. If you think about the credit card business and the way it was established, you had the traditional players with a single, ‘one size fits all' product and many of them charging huge interest rates. Until recently, you either went to Barclays and got a Barclaycard or NatWest and got an Access card. It was the same in the US too, where Capital One started up.
The traditional players have not touched people in the higher credit risk bracket with a barge pole. As a general rule, they target the low-risk end of the market. The philosophy within the Capital One environment is that, actually, if you understand your data properly, you will find that even in the high-risk bracket, most people do actually pay back and you can still make a healthy profit. The trick is working out who won't. We have some of the lowest rates of default in the industry, so we do know how to do it.
At Capital One we have deployed an information-based strategy, which is based around getting customer data from whatever sources are available, loading it onto our systems and analysing the hell out of it. This means we can put together products that will appeal to different types of people, by testing whether hypotheses are right, looking at the result of these tests, modifying the hypotheses and testing again, and so on.
IA: Given the stress you put on data, this must place a huge strain on Capital One in terms of data management. What kind of data warehousing and data mining infrastructure do you have in place to support your information-based strategy?
CD: I can't remember how many terabytes we have exactly, but I do know that it's the equivalent of 354 million telephone directories. We have a huge machine in the US and a huge machine in the UK and one of the things we do is share data between the two. It's just vast.
IA: With this immense amount of information at your fingertips, analysing it must be an enormous challenge. Where does Capital One source its data on existing and potential customers, and how do you go about analysing it?
CD: We're data hungry beasts and try to glean data from as many different sources as we can, providing it's legal and ethical. So we get data from standard places such as the Post Office address file and credit checking agencies such as Equifax and Experian. But we offer a range of ‘lifestyle' products, such as a gardening credit card, so we get data from organisations that might target people with such interests.
We bring all the data in, we normalise it and clean it up using a software package from Trillium. And then we use tools from business intelligence software vendors such as SAS Institute and Brio for deep analysis. You end up with a body of knowledge that has accumulated over time. Soon you begin to see that the behaviour people exhibit indicates whether they are more or less likely to default.
For instance, if you offer someone payment protection insurance, their response to that offer is an indicator as to whether you needed to offer it to them in the first place. Some of these indicators are counter-intuitive, so when a result turns up that you're not expecting, it's really very interesting. It's all about the ‘Ah-hahs' in life.
IA: How have you turned the success you've had in analysing and using data into better customer service?
CD: The main benefit for customers is that that people that otherwise would not find it easy to get credit can get credit from us. Although some of the more traditional players are starting to dip their toe into the higher credit risk bracket, they've only done so very slightly. At the end of the day though, we're not a charity, we're a business. Part of what we do is to tailor the products and the charges associated with them relative to the profile of the various customer segments. The whole purpose of our data analysis machine is that we want to identify upfront who are the people that are going to pay us back and who are not.
IA: This technology-centric, information-based strategy places a lot of focus on your position as CIO. How does Capital One bridge the traditional divide between the IT and business functions, and how does this affect your role on a day-to-day basis?
CD: It is certainly true to say that technology is a core component of how we do our business. In more traditional or historical organisations, the IT guys tend to have a sense of inferiority, but that doesn't happen here. Before I joined Capital One I actually had a conversation with one of the founders, Nigel Morris, about the relationship between business and IT. Having worked in some of the larger institutions, I knew what different types of relationships there could be. Morris said that in Capital One, he thought the relationship [between the IT and business functions] was the best he'd ever seen, but it still wasn't where he wanted it to be.
So when we have discussions about new strategies and products, both operations and IT are involved in those discussions rather than us getting a letter on Monday morning saying, ‘Can you implement this on Wednesday?' That's not to say it's perfect. Sometimes if someone's got a bright idea they forget to think about some of the IT logistics. But that's the exception. The norm is that we get involved.